Hypothetical examples of dental clinical trial data were developed to demonstrate the utility of a statistical technique of blocking on baseline DMFS. In the five trials studied, crude means ranged from one extreme of highly significant caries reduction in the fluoride group for one trial to the opposite extreme of highly significant caries reduction in the control group for another trial. Blocking on baseline DMFS, however, revealed that the trials were essentially identical.